From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 10094 invoked by alias); 23 Mar 2010 20:57:09 -0000 Received: (qmail 10081 invoked by uid 22791); 23 Mar 2010 20:57:07 -0000 X-SWARE-Spam-Status: No, hits=-7.3 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_HI,SPF_HELO_PASS X-Spam-Check-By: sourceware.org Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 23 Mar 2010 20:57:01 +0000 Received: from int-mx04.intmail.prod.int.phx2.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.17]) by mx1.redhat.com (8.13.8/8.13.8) with ESMTP id o2NKuxf1001719 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Tue, 23 Mar 2010 16:57:00 -0400 Received: from host0.dyn.jankratochvil.net (ovpn01.gateway.prod.ext.phx2.redhat.com [10.5.9.1]) by int-mx04.intmail.prod.int.phx2.redhat.com (8.13.8/8.13.8) with ESMTP id o2NKuvQ2020614 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Tue, 23 Mar 2010 16:56:59 -0400 Received: from host0.dyn.jankratochvil.net (localhost [127.0.0.1]) by host0.dyn.jankratochvil.net (8.14.3/8.14.3) with ESMTP id o2NKuumv012824 for ; Tue, 23 Mar 2010 21:56:56 +0100 Received: (from jkratoch@localhost) by host0.dyn.jankratochvil.net (8.14.3/8.14.3/Submit) id o2NKuu8U012823 for gdb-patches@sourceware.org; Tue, 23 Mar 2010 21:56:56 +0100 Date: Tue, 23 Mar 2010 20:57:00 -0000 From: Jan Kratochvil To: gdb-patches@sourceware.org Subject: [patch] Fix separate-debug with non-unique section names (PR 11409) Message-ID: <20100323205655.GA12124@host0.dyn.jankratochvil.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.20 (2009-08-17) X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2010-03/txt/msg00799.txt.bz2 Hi, gdb-7.1 is now broken for example for debugging /usr/bin/emacs due to: http://sourceware.org/bugzilla/show_bug.cgi?id=11409 [22] .data PROGBITS 00000000007fe8a0 1fe8a0 215068 00 WA 0 0 32 [23] .data PROGBITS 0000000000a13920 413920 68c6e0 00 WA 0 0 32 It is in fact a regression against gdb-7.0 by me due to: commit 71d0069a9f238a11f7f455bf6ad2adfc25683521 Author: Jan Kratochvil Date: Tue Jan 5 15:51:01 2010 +0000 gdb/ * symfile.c (syms_from_objfile): Remove the !MAINLINE conditional. as while the code was broken even before the broken relocation was not applied to mainline binary (before PIE+OSX patches went in). No regressions on {x86_64,x86_64-m32,i686}-fedora12-linux-gnu. OK to check-in also for gdb-7.1 (7.1.1)? Thanks, Jan gdb/ 2010-03-23 Jan Kratochvil * symfile.c (addr_info_make_relative): Move sect declaration to the outer block. Initialize it to NULL. Prefer SECT->next more than bfd_get_section_by_name. gdb/testsuite/ 2010-03-23 Jan Kratochvil * gdb.base/dup-sect.exp, gdb.base/dup-sect.S: New. --- a/gdb/symfile.c +++ b/gdb/symfile.c @@ -529,6 +529,7 @@ addr_info_make_relative (struct section_addr_info *addrs, bfd *abfd) asection *lower_sect; CORE_ADDR lower_offset; int i; + asection *sect; /* Find lowest loadable section to be used as starting point for continguous sections. */ @@ -553,11 +554,23 @@ addr_info_make_relative (struct section_addr_info *addrs, bfd *abfd) (the loadable section directly below it in memory). this_offset = lower_offset = lower_addr - lower_orig_addr */ + sect = NULL; for (i = 0; i < addrs->num_sections && addrs->other[i].name; i++) { const char *sect_name = addrs->other[i].name; - asection *sect = bfd_get_section_by_name (abfd, sect_name); + /* Prefer the next section of that we have found last. The separate + debug info files have either the same section layout or just a few + sections are missing there. On the other hand the section name is not + unique and we could find an inappropraite section by its name. */ + + if (sect) + sect = sect->next; + if (sect && strcmp (sect_name, bfd_get_section_name (abfd, sect)) != 0) + sect = NULL; + + if (sect == NULL) + sect = bfd_get_section_by_name (abfd, sect_name); if (sect) { /* This is the index used by BFD. */ --- /dev/null +++ b/gdb/testsuite/gdb.base/dup-sect.S @@ -0,0 +1,22 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2010 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + + .section sect1, "a" +var1: .byte 1 + + .section sect2, "a" +var2: .byte 2 --- /dev/null +++ b/gdb/testsuite/gdb.base/dup-sect.exp @@ -0,0 +1,79 @@ +# This testcase is part of GDB, the GNU debugger. + +# Copyright 2010 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +# Test inappropriate offseting of multiple sections with the same name. +# When kept in object file (before final executable link) it still works. +# When separate debug info file is not used it still works. +# When the ELF symbol table is kept in the main binary it still works. +# Used .S file as in .c file we would need __attriute__((section)) which is +# a GCC extension. + +# This test can only be run on targets which support ELF and use gas. +# For now pick a sampling of likely targets. +if {![istarget *-*-linux*] + && ![istarget *-*-gnu*] + && ![istarget *-*-elf*] + && ![istarget arm-*-eabi*] + && ![istarget powerpc-*-eabi*]} { + return 0 +} + +set testfile dup-sect +set srcfile ${testfile}.S +set srcmainfile start.c +set executable ${testfile} +set binfile ${objdir}/${subdir}/${executable} + +if {[build_executable ${testfile}.exp $executable [list ${srcfile} ${srcmainfile}] {}] == -1} { + return -1 +} + +set test "rename section" +set objcopy_program [transform objcopy] +set result [catch "exec $objcopy_program --rename-section sect2=sect1 $binfile" output] +verbose "result is $result" +verbose "output is $output" +if {$result != 0} { + fail $test + return +} +pass $test + +set test "split" +if {[gdb_gnu_strip_debug $binfile] != 0} { + fail $test +} else { + pass $test +} + +# gdb_gnu_strip_debug uses only --strip-debug and keeps the ELF symbol table +# in $binfile. +set test "strip" +set strip_program [transform strip] +set result [catch "exec $strip_program $binfile" output] +verbose "result is $result" +verbose "output is $output" +if {$result != 0} { + fail $test + return +} +pass $test + +clean_restart $executable + +gdb_test "p/d *(const char *) &var1" " = 1" "var1 after strip" +gdb_test "p/d *(const char *) &var2" " = 2" "var2 after strip"